Daniel Szmulewicz wrote:
> Sebastien Vauban wrote:
>> AFAICT, when using the `utf8' option to `inputenc' (and not `utf8x'
>> which should be avoided, as it uses `ucs' which is no longer
>> maintained), we have 2 solutions to support the UTF-8 non-breaking
>> space:
>>
>> - Convert it in Org mode, when exporting (via a filter)
>>
>> #+begin_src emacs-lisp
>> ;; filter for non-breaking spaces
>> (defun leuven--latex-filter-nbsp (text backend info)
>> "Convert non-breaking spaces when exporting to LaTeX/Beamer."
>> (when (memq backend '(latex beamer))
>> (replace-regexp-in-string "?" "~" text)))
>>
>> (add-to-list 'org-export-filter-plain-text-functions
>> 'leuven--latex-filter-nbsp)
>> #+end_src
>>
>> - or Convert it in LaTeX:
>>
>> #+begin_src emacs-lisp :tangle no
>> ;; convert `nbsp' to its LaTeX equivalent
>> (add-to-list 'org-latex-packages-alist
>> (concat "\\ifdefined\\DeclareUnicodeCharacter{"
>> "\\DeclareUnicodeCharacter{00A0}{~}"
>> "}\\fi") t)
>> #+end_src
>>
>> I think that the first one is "better", though its problem is that,
>> being done by a filter (hence, on the user-side), sharing an Org file is
>> not easy anymore (without sharing the filter to be added to the
>> personal .emacs file).
>
> Awesome. Thank you, Sebastien.
>
> I prefer the second solution, because it allows me to account for
> different kinds of spaces: espace fine, espace insécable, espace
> moyenne, etc.
>
> Just for curiosity: I’m not sure what to make of the regexp. The ‘?’
> is normally a postfix operator used in conjunction with other
> characters, isn’t it? As it is, it seems to target question marks
> instead of blank spaces. What am I missing?
It seems to be a character encoding problem: it was supposed to be
a real "nbsp" char (00A0, IIRC).
It's clearly NOT intended to be a "?". I don't know how this got badly
converted in these posts.
Best regards,
Seb
--
Sebastien Vauban
